From: Bona Craig (bona_64unix@yahoo.com)
Date: Tue Jul 15 2003 - 07:40:21 EDT
Hi all,
I have two ES40s running in a cluster. The cluster is
connected to a SAN (HSGs, controlling 3 EMA 12000
each).
| Omega |________________ |Midas |
| | | |
----------- -----------
/alpha
/beta
•Omega is serving the IO for the filesystems mounted
on /alpha and /beta.
•A copy process is initiated on Midas to copy a file
from /beta to /alpha.
•The file size is 10GB in size.
•We can see that all of the Read IO for /beta is being
served by Midas.
•We can see all of the write IO for /alpha is being
served by Omega.
•If we run the command “# ls –ltr /alpha” on Midas
then the command does not return until one of the
following conditions:
oThe copy command is completed
oThe copy command is aborted
oWait 3 - 10 minutes
•If we run the command “# ls /alpha” on Midas the
command returns immediately
•If we initiate the copy from Omega then the “# ls
–ltr /alpha” command responds immediately.
Further observations.
•In the same scenario where Omega is serving the IO
•A copy is initiated on Omega from /beta to /alpha
•Go to the /alpha directory
•On Omega ls –ltr takes .001 seconds - 1.0 seconds
•On Midas ls –ltr takes 3 - 10 Seconds
•If You initiate an ls –ltr on Midas and then
immediately on Omega. The ls –ltr on Omega only
returns after the command on Midas has completed
Does anyone have ANY idea what could be causing this?
Thanks
__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:49:27 EDT